Reversible stochastic epigenetic like silencing of the production of pulcherriminic acid in the antimicrobial antagonist Metschnikowia Pulcherrima Authors: Matthias Sipiczki, Kinga Czentye Published: 2024 ABSTRACT AbstractThe ability to produce pulcherriminic acid is a characteristic feature of yeast species of the pulcherrima clade recently merged under the taxonomic name Metschnikowia pulcherrima. This iron chelator cyclodipeptide forms pulcherrimin, a maroon-red pigment with ferric ions. Its synthesis and secretion into the environment is under the control of closely linked genes referred to as the PUL cluster. The examination of 18 generations of single-cell clones generated from a stock culture of the collection strain 11-1090 (CBS 10359) in this study revealed that the biosynthesis of pulcherriminic acid is reversibly switched on and off during the propagation of cells in a way similar to the epigenetic silencing and activation of gene expression (bimodal active/silent state) in near-heterochromatic regions of other yeast species. As the strain is heterozygous for PUL2 alleles encoding slightly different amino acid sequences and has a plastic genome structure, the efficiency of pulcherriminic acid synthesis in the switched-on state is presumed to depend on which PUL2 allele is active and on structural changes in the genome. The transitions between the active and silent states of pulcherriminic acid synthesis are associated with transitions between the active and silent states of antimicrobial antagonism. This association confirms the primary role of pulcherriminic acid in the antimicrobial antagonism of M. pulcherrima.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-024-80436-9.Keywords: Antagonistic yeast, Pulcherrimin, Variegation, Segregation, PUL clusterSubject terms: Genetics, Microbiology Metschnikowia pulcherrima (synonyms: Torula pulcherrima, Torulopsis pulcherrima, Candida pulcherrima) is an ascomycetous yeast species1whose colonies can produce pulcherrimin, a maroon-red pigment on media containing ferric ions2–5. Over the past twenty years numerous other pulcherrimin-producing Metschnikowiaspecies have been described6,7. Because of their close phylogenetic relationship with each other and with M. pulcherrima, the oldest known pulcherrimin-producer of the genus, these species were frequently referred to as the pulcherrima clade (e.g6). The clade comprised eight species (M. andauensis, M. fructicola. M. leonuri, M. pulcherrima, M. rubicola, M shanxiensis, M. sinensis and M. zizyphicola) and two groups of strains whose taxonomic names (M. citriensis and M. persimmonesis) are listed in Mycobank as invalid (“nomen invalidum“ referring to a name not published in accordance with rules specified in the International Code of Nomenclature for algae, fungi, and plants8). However, recent analyses of primary and secondary barcode sequences of ex-type cultures and their mating tests revealed that the species of the clade are neither separated by barcode gaps nor isolated by reproductive barriers9. Consequently, the pulcherrima-clade strains were merged in one species under the oldest species name, M. pulcherrima10. A common, species-specific property of the strains of the merged species is the ability to produce pulcherrimin. This pigment is a complex of pulcherriminic acid and ferric ions5,11. The complex-forming reaction can take place both within the yeast cells and extracellularly because the pulcherriminic acid can be released into the environment. The intracellular production of pulcherrimin turns the colony red, whereas the extracellularly produced pulcherrimin forms a pigmented halo around the colony. The biological role of pulcherrimin production is obscure and subject of debate. Kupfer et al.12assumed that the formation of pulcherrimin is a fortuitous event depending on the availability of iron in the environment, and has no effect on cell metabolism. According to Beijerinck’s hypothesis3its production is a defence reaction of the organism against a harmful effect of an excess of iron present in the medium. The alternative hypothesis is that it improves the competitiveness of the producer by making the iron inaccessible to other microbes13. Since iron is essential for the activity of many proteins and cellular processes, its immobilisation by pulcherriminic acid adversely affects the propagation of other microorganisms and even kills their cells. Due to their antimicrobial activities, the pulcherrimin-producing yeasts can be used as biological agents to protect agricultural commodities and food products against pathogenic and destructive microorganisms (for a review, see14). As the reaction of pulcherriminic acid with ferric ions is irreversible under conditions suitable for the growth of yeast cells and the resulting pulcherrimin is almost completely insoluble in water3,15, the extracellular production of pulcherrimin is unlikely to play a role in iron acquisition. By analogy with its production in Bacillus and Kluyveromyces lactisstrains for which molecular data are available16–18, the pulcherrimin of M. pulcherrima is the end product of a biosynthetic pathway consisting of two enzymatic and one spontaneous (non-enzymatic) reaction. The enzymes are encoded by two closely linked genes (PUL1 and PUL2) that form a group (referred to as the PUL cluster) with the gene (PUL4) that codes for their regulator and a gene (PUL3) of uncertain function9,16. An interesting feature of pulcherrimin production in M. pulcherrimais that the intensity of pigmentation frequently varies both among strains and within strains (e.g3,4). Punkari and Henrici4 noted colour variants even within colonies appearing as sectors or papillae. Many of them were subcultured but no stable variants could be obtained. The authors hypothesized that the formation of sectors and papillae might be due to accidental “mutation-like changes”. Kluyver et al.5believed that the pigment production was a property of a mutant form (“red mutant“). Consistent with these early observations, numerous recent studies reported on intraspecies variability of the production of pulcherrimin and the efficiency of antimicrobial activity of strains (e.g19–23). In the diversity study of a large number of wild isolates, it was noticed that almost all isolates had non-pigmented sectors in their colonies24. The observed instability of pigmentation indicates that the production of pulcherriminic acid is a variable property of the species and the cultures consist of different subpopulations. It can be assumed that the presence of non-productive (silent) subpopulations impairs the antagonistic efficacy of the culture. The more frequent the switch to the silent state, the less effective the bioprotective activity of the culture. It was further hypothesised that epigenetic changes may be involved in the switches24. The aim of the present study was to investigate the switches (transitions) between the active and silent states by monitoring the changes of colony pigmentation. We examined six ex-type strains of the merged species to see if they also tended to segregate. All turned out to have unstable pigmentation phenotypes, albeit with very different segregation frequencies. The changes of the pigmentation phenotypes of 18 generations of clones arising from single cells of the previous generation indicate that the transitions (switching) between the active and silent states of the production of pulcherriminic acid and the associated antimicrobial activity are controlled by processes similar to the epigenetic silencing and activation (stochastic active/silent mode) of gene expression in near-heterochromatic regions of other yeast species (these epigenetic mechanisms are described in Discussion). The intensity of pigmentation in the switched-on state can be affected by the number and versions of the activated PUL clusters and/or by changes of the genome structure. Results Pulcherrimin production is unstable in ex-type cultures of former pulcherrima-clade species The ex-type cultures of 6 validly described species of the pulcherrima clade were tested for the stability of pulcherrimin production on agar plates supplemented with FeCl3. In the tests, the cultures formed both pigmented and pigment-free and/or less intensely coloured colonies (Fig. 1) or sectors and patches (Fig. 2). Upon longer incubation allowing the growth of the sectors, the larger sectors often segregated to produce pigmented “subsectors”. The least unstable strain was 11–1120 (type of the former species M. andauensis) and the most unstable strain was 11–1090 (type of the former species M. shanxiensis) (Fig. 1). The latter was chosen for the investigation of the changes of pigmentation during vegetative growth. All other results presented in this study relate to this strain. Segregation of pulcherrimin production. (a) 11–578. (b) 11–579. (c) 11–1088. (d) 11–1089. (e) 11–1090. (f) 11–1120. Segregation of pulcherrimin production in strain 11–1090. (a) Differently pigmented sectors at the edges of a culture formed by cells smeared on the surface of the medium. (b) A culture formed by a cell suspension of the same strains dropped on the medium. (c) The transilluminated image of (b). Intraclonal variation (segregation) of pulcherrimin production To study the changes of pulcherrimin production, 11–1090 cells were spread onto iron-containing plates to obtain detached colonies. Since the colonies arose from single cells, each was considered an individual clone of vegetatively propagating isogenic cells possessing complete PUL clusters. It was assumed that if pulcherrimin synthesis was switched on (active PUL cluster) in the founder cell, a homogeneously pigmented colony was produced. If the PUL cluster was silent in the founder cell, the colony was presumed to be pigment-free. In the case of changes of the activity of PUL genes during the growth of the colony (propagation of cells), sectors and/or patches of different pigmentation were expected within the colony. Both homogeneous and sectored colonies were formed on the plates (Fig. 3a and b). Diverse types of colony pigmentation and shape. (a) Various degrees of homogeneous pigmentation. (b) Colonies with differently pigmented sectors and patches. The second photograph shows a transilluminated image. (c) Diversity in colony shape and ornamentation. Two homogeneously dark red and two homogeneously non-pigmented colonies (clones) were selected to produce the second generation of clones. The colonies formed by the cells of a dark red and a non-pigmented first-generation clone were homogeneously pigmented and non-pigmented, respectively. The cells of the other two clones produced mixed populations of colonies (Fig. 4). Thus, even clones looking phenotypically homogeneous could contain cells in which the activity of the PUL genes had changed. Pedigree of eighteen generations of clones. One sectored second-generation colony was then selected for further investigation. Its cells were spread onto fresh plates to produce another (third) generation of clones. The colonies obtained had diverse pigmentation morphologies: homogeneously dark, homogeneously non-pigmented and variously sectored (Fig. 4). From one colony of each type, cells were spread onto fresh plates to obtain the fourth generation of clones. Two of them proved to be mixed. The sectored four-generation colonies produced by a non-pigmented third-generation clone were used to produce the fifth generation of clones. All produced mixed populations of colonies. The procedure was then repeated until the 18th generation of clones was obtained (Fig. 4). The overwhelming majority of clones of all types of pigmentation proved to be unstable. In the eighth generation an additional phenotype appeared: faint pigmentation (brownish colonies). The cells of the faintly pigmented clones also often produced heterogeneous populations of colonies including dark red, pigment-free and sectored colonies. From the recurrent loss and regain of the ability to produce pulcherrimin in the 18 generations of clones, it can be inferred that the biosynthesis of pulcherriminic acid is reversibly silenced and reactivated during cell propagation. Intraclonal heterogeneity of colony size and morphology In addition to being diverse and unstable in pigmentation, the clones also varied in other morphological characteristics. Examples differing from the “standard” morphology (shown in Fig. 3a) are presented in Fig. 3c. Differences were found in surface ornamentation (smooth, rough, wrinkled, glistening, dull) and margin shape (entire, undulate) and also in size (Fig. 5), Like pigmentation, the colony morphology types also frequently segregated (e.g. Figure 3c) but their segregation did not correlate with the segregation of pigmentation. Segregation of colony size in the 9th generation of clones. (a) Clone 3/3. (b) Clone 6/2. (c) Clone 1/1. Diverse RAPD patterns indicate structural plasticity of genomes To examine whether the differences in pigmentation were associated with structural changes in the genomes, all clones selected from the ninth generation for segregation tests (Fig. 4) and the stock culture of 11–1090 were subjected to RAPD analysis with five primers. All primers revealed differences but no single primer produced unique patterns for each clone (Table 1; Fig. 1S). Nevertheless, when taking all patterns of all primers together, each clone differed from all other clones and also from the stock culture of 11–1090. Remarkably, certain bands present in multiple patterns varied in brightness. Table 1RAPD analysis of clones of generation 9. 1for symbols of pigmentation, see Fig. 4. 2 the presence of bands is marked with blue colour. Light blue: faint band. RAPD analysis of clones of generation 9. 1for symbols of pigmentation, see Fig. 4. 2 the presence of bands is marked with blue colour. Light blue: faint band. PUL2 heterozygosity The genome of the strain 11–1090has been sequenced25 and the assembly ASM3057869v1 is now available in GenBank. It has one set of PUL genes in scaffold Sc0000111. Contrary to this, the strain proved to be heterozygous for PUL2 in this study. The gene PUL2encodes a putative cytochrome P450 monooxygenase which is thought to convert the precursor (cyclic dipeptide cyclo(L-leucyl-L-leucyl), into pulcherriminic acid16,26. Here an internal segment of PUL2 was amplified from 11 to 1090 and the amplicon was sequenced. The sequence matched perfectly its counterpart in the genome sequence except for 8 ambiguous nucleotides (Table 2). All ambiguous sites were dimorphic (SND sites; SND: single nucleotide dimorphism). The alternating nucleotide pairs were C/T (Y) in five sites and A/G (R) in three sites. Their ambiguity might be attributed either to sequencing errors or to the presence of different PUL2 alleles in the genome. As re-sequencing gave rise to the same sequence, the former possibility was unlikely. Table 2Nucleotide substitutions in the PUL2 genes of 11–1090. WGS: whole genome sequencing. 1The numbers represent the locations of the nucleotides in the coding region.Strain/cloneMode of sequencingNucleotides in dimorphic sites of 11–1090 (marked with colour) and their position in amino acid codons1889912997101110631158120012091225APC 1.2.WGSACAGAGACACTCGGCAAGGGTTTCTTC11–1090WGSACGGAATCACTTGGCAGGGGCTTTTTCampliconACRGARTCACTYGGYARGGGYTTYTTY1clonedACAGAGACACTCGGTAAGGGTTTCTTT2clonedACAGAGACACTCGGTAAGGGTTTCTTT3clonedACAGAGTCACTTGGCAGGGGTTTCTTTAclonedACAGAGTCACTCGGTAAGGGTTTCTTT4clonedACAGAGACACTCGGTAAGGGTTTCTTT5clonedACGGAATCACTTGGCAGGGGCTTTTTC6clonedACGGAATCACTTGGCAAGGGTTTCTTTBclonedACGGAATCACTTGGCAGGGGCTTTTTCCclonedACGGAATCACTTGGCAGGGGCTTTTTCDclonedACGGAATCACTTGGCAGGGGTTTCTTTEclonedACGGAATCACTTGGCAGGGGCTTTTTCTranslationThrGluSerorThrLeuGlyArgorLysGlyPhePhe Nucleotide substitutions in the PUL2 genes of 11–1090. WGS: whole genome sequencing. 1The numbers represent the locations of the nucleotides in the coding region. Ser or Thr Arg or Lys To demonstrate that the genome was heterozygous for PUL2 alleles, individual DNA molecules were cloned from the amplified PUL2 DNA and sequenced. None of the eleven cloned sequences contained ambiguous nucleotides. At the positions corresponding to the SND sites of the amplicon sequence, the clones had one or the other nucleotide of the alternating pairs. Unexpectedly, an R-type transition was also found in a site which was not dimorphic in the amplicon sequence. Two sequences were obtained two times and one sequence was obtained three times, indicating that the genome had at least three different versions of the PUL2 gene. Remarkably, none of the cloned sequences was identical to the sequence of the PUL2 gene of the genome sequence available in GenBank (Table 2). The possible cause of this discordance is that the GenBank sequence is a haploid assembly, in which all copies of homologous gene sequences are collapsed into single consensus sequences. Thus, the PUL2 sequence of this assembly can be an artefact (created by the assembling algorithm) that does not exist in the cells. None of the cloned sequences was identical to the corresponding segment of the PUL2 gene of the genome sequence of APC 1.2. (assembly accession number ASM421770v1), the only M. pulcherrimagenome sequence assembled to chromosomes26, either. Seven dimorphic sites were in third positions of codons in which the nucleotide substitutions did not cause amino-acid differences between the translated sequences (Table 2). One site was in the first and another site in the second position of codons and both caused amino acid substitutions, demonstrating that different Pul2 proteins existed in the cells. The sequence of scaffold Sc0000111 showed 96% identity to the segment located between nucleotides 1,421078 and 1,446,431 of chromosome III (total assembly length: 2,688,662 nucleotides) of the APC 1.2 genome. Active and silent states of pulcherriminic acid synthesis correlate with active and inactive antimicrobial antagonism All clones of generation 9 were subjected to antagonism tests against two indicator organisms. The heavily pigmented clones (3/1, 3/2, 4/1, 6/3) and the clones with dark segments (3/3, 6/2) formed inhibition zones in the lawns of both indicator organisms. The non-pigmented clones (1/2, 4/2, 6/2) and the brownish clones (1/1, 2/1, 5/1, 5/2) had no inhibitory effect detectable by this test method. Figure 6 shows examples. Despite the fact that the medium was not supplemented with iron, faint pigment rings were visible on the edges of the inhibition zones. Antagonism test on lawns of indicator organisms. Lawn is Zygosaccharomyces aff. siamensis 11–2106 in (a) and Debaryomyces aff. hansenii Vm7 in (b). The tested clones are the brown clone 5/1 (a), the non-pigmented clone 1/1 (b), the heavily pigmented clone 3/2 (c) and the sectored clone 3/1 (d) from the ninth generation of clones (see Fig. 4). Red arrow: faint pigment ring on the edge of the inhibition zone. Blue arrow: weak inhibition around the brown clone. Discussion Pulcherrimin production proved to be unstable in all ex-type strains of the former species of the pulcherrima clade of M. pulcherrima involved in this study. The pigmented cultures growing on agar plates frequently formed non-pigmented sectors on their edges or patches on their surfaces. The sectors and patches also turned out to be unstable. In most of them, pigmented subsectors appeared during longer periods of growth. The instability of both the pigmented and the non-pigmented forms, as well as their reversible conversions, indicate that neither form can become exclusive and therefore both pigment-producing and pigment-free populations can persist in the cultures even if they may differ in fitness. Consistent with this conclusion, the cells of 11–1090, the least stable strain, spread onto iron-containing agar plates formed both pigmented and pigment-free colonies. Since the pigment is produced by pulcherriminic acid, the formation of these colonies demonstrates that the 11–1090 culture maintained in the laboratory consisted of cells in which the biosynthetic pathway of pulcherriminic acid was active and cells in which the pathway was silent. As the colonies were established by single cells, they were considered clones of vegetatively propagating cells. When cells of pigment-free clones were spread on fresh plates, predominantly pigment-free colonies (clones) were formed but some contained pigmented sectors and/or patches (segregating clones). One of the segregating clones was selected for the generation of a series of descendant clones. The clones of the 16 subsequent generations could be grouped in the same basic pigmentation categories (pigmented, pigment-free and segregating) although the degree of pigmentation varied in the pulcherrimin-producing clones and sectors. When all clones of all generations were considered, it could be concluded that the production of pulcherriminic acid was reversibly silenced and activated during the vegetative propagation of cells. The formation of distinct sectors and patches further indicated that the switches did not occur at each cell division but stochastically over several generations of cells. Cells of many organisms are capable of expressing multiple phenotypes from a single genome when exposed to different environmental conditions. The genetic programmes needed to produce alternative phenotypes are encoded within the genome, and environmental stimuli determine which programme is expressed. These phenotypic changes are essential for the adaptation of the microorganisms to changes in the environment and for the development of organs in multicellular plants and animals. In contrast to these sorts of reprogramming of gene expressions, the transitions between the observed silent and active state of pigmentation (pulcherriminic acid synthesis) in M. pulcherrima were not elicited by environmental stimuli because the same medium and culturing conditions were applied throughout the stability tests and the process of isolation and maintenance of the clone generations. Certain earlier studies considered the pigmented form a mutant version of the pigment-free wild-type form (e.g5). If the non-producer-to-producer transition is caused by a mutation, the reverse transition must occur by a reverse mutation. However, most genetic changes have negligible spontaneous reversion rates27. Therefore, it is unlikely that forward and backward mutations underline the transitions between the active (producing) and silent (non-producing) states. The situation is further complicated by the presence of more than one PUL cluster in the same genome. A search for PUL genes in genome sequences identified two PUL clusters in certain M. pulcherrimastrains9. A recent study revealed PUL4heterozygosity in a large group of isolates24. PUL4 is also part of the PUL cluster. The cloning of PUL2 in this study identified PUL2 heterozygosity in 11–1090. As the four PUL genes are closely linked in all sequenced M. pulcherrima genomes, including that of 11–1090, this result indicates that 11–1090 has more than one PUL cluster. Thus, a fully OFF (pigment-free) state would require “inactivation mutations” in all copies of the PUL clusters and the return to the dark-pigmentation phenotype would require back mutations in all clusters. The probability of such coincident mutational events in different loci is very low (the likelihood of concomitant spontaneous mutations in two clusters is equal to the square of the likelihood of one spontaneous mutation in one cluster). Therefore the transitions between the pigmented and non-pigmented phenotypes are unlikely to be caused by mutations. We therefore hypothesized that the transitions were due to changes in the level of transcription. Instead of being inactivated by a mutation, the cluster can be reversibly silenced in a way similar to the telomeric position effect (TPE) variegation described in Saccharomyces cerevisiae28. When the ADE2 gene was inserted into a telomere boundary region of S. cerevisiae, the cells produced colonies that were predominantly either white (ADE2 expressing) with red (ADE2 repressed) sectors, or red with white sectors. The silent state of the ADE2 gene was stably passed on to following cell generations but it was reversible. The ON state was also heritable and reversible. As no sequence differences were detected between the ON and OFF states of ADE2, the phenomenon was attributed to epigenetic control of gene expression29. A similar phenotype was also observed when ADE2 was inserted into the HMR-right region30. In Schizosaccharomyces pombetranscription silencing occurs in pericentromeric regions31. The mechanisms underlying the transitions between the silent and expressed states in these examples are thought to be changes in heterochromatin-like structures (cis-spreading of a condensed heterochromatic state) and histone modifications near the inserted genes (for reviews, see32,33). Single cell analysis in S. cerevisiaeshowed that the establishment of telomeric silencing occurs stochastically34. The similar stochastic ON-OFF transitions in the production of pulcherrimin (pulcherriminic acid) might be due to a similar epigenetic mechanism. Unfortunately, the location of the PUL cluster in the 11–1090 genome is not yet known because the genome sequence available for this strain is only assembled to scaffolds. In the only chromosome-level assembly of the species (strain APC1.2; accession number ASM421770v1) the PUL cluster is located roughly halfway between the telomeres of chromosome III. The position of the centromere of the chromosome is not defined in the assembly. However, since the genome structures of the M. pulcherrimastrains are variable9,10, the location of the cluster may be different in 11–1090 and in APC 1.2. Even if it is not adjacent to any heterochromatic region, it still may come under their influence by “trans-interaction” with them in a way similar to that observed in Drosophila. In Drosophila, interactions via DNA loop formation were detected between distantly located euchromatin and heterochromatin so that the heterochromatin became capable of quenching the expression of euchromatin35. This mechanism is assumed to operate in plants as well36 and may also exist in Metschnikowia. Another possibility is trans-regulation, in which the changes in a distantly located heterochromatic region indirectly modulate the activity of the PUL cluster, via silencing and reactivation of their regulator gene(s) such as SNF2whose overexpression was found to reactivate pigment production in a non-pigmented mutant26. In addition to the position effect variegation, position-independent epigenetic switching was also described in S. cerevisiae(for recent reviews, see37,38). For example, the expression of certain FLO and PHO genes positioned away from the telomeres were found to switch reversibly between heritable ON and OFF states by promoter-specific mechanisms involving the HDA1 histone deacetylase complex. However, the switching of the transcription of these genes is controlled by environmental stimuli (e.g. nitrogen starvation in the case of FLO11). In strains of M. reukaufii, a species belonging to a different clade of the genus39, a different type of epigenetic regulation of gene expression was observed. Herrera et al.40 found that DNA methylation profiles changed in the cells of these strains following exposure to various sugar-containing media. Those changes were induced by environmental stimuli (e.g. changes in sugar concentration), whereas the transitions between the ON and OFF states of pulcherrimin (pulcherriminic acid) production occurred in this study under static culturing conditions. It is also worth considering that DNA CpG methylation is not ubiquitous in yeasts. It is absent in Saccharomyces, Schizosaccharomyces and Komagataella (formerly Pichia) pastoris. These organisms rely primarily on histone modification to change epigenetic states41. Due to their ability to release pulcherriminic acid in the environment, M. pulcherrimastrains can inhibit the growth of other microorganisms13,42,43. The direct connection between pulcherriminic acid production and antimicrobial antagonism is corroborated by the antagonism tests performed in this study. The heavily pigmented clones formed broad inhibition zones in the lawns of the indicator microorganisms whereas the non-pigmented clones had no effect on the lawn. However, extensive synthesis of pulcherriminic acid can be a burden to the Metschnikowiacells because it takes resources from other metabolic processes and can interfere with iron homeostasis even in the producer cells. In a study by Roberts44 all strongly pigmented cells proved to be non-viable. In general, should any phenotype have selective advantage on the laboratory medium, it should have outcompeted the other phenotypes during the years of maintenance of the 11–1090 culture that have elapsed since the culture was obtained from the CBS collection. Obviously, this has not happened. The non-producer version has not outcompeted the producer version. In this study, there were clones in which the pigmented sectors and patches grew faster than the non-pigmented parts of the colony but in other clones the opposite was observed. Another argument against a significant adverse effect of pulcherrimin production on the growth of the producer was the simultaneous formation of larger and smaller pigmented colonies by certain clones and/or the formation of small non-pigmented colonies among larger pigmented colonies by other clones. Apparently, the colony size (growth efficiency) was more affected by other factors. Remarkably, the colonies did not vary only in pigmentation and size. Differences were also seen in surface ornamentation and in the shape of colony edges. Although the inheritance of these morphological properties was not examined in this study, it is tempting to suppose that structural changes in the genomes may be behind these changes. This assumption is justified by observations made on other yeast species and by the variable copy numbers of certain genes in M. pulcherrimagenomes. In numerous yeasts the changes of ploidy (e.g. aneuploidisation) can cause changes in colony morphology (e.g45–49). As for M. pulcherrima, recent studies revealed genomic mosaicism and different copy numbers of the PULcluster and certain other genes in its strains9,10. The genome size of the sequenced strains of the species varies between 15,5 and 27.5 Mb (GenBank; 5.10.2024). In this study the genomes of a group of clones were compared by RAPD analysis. Each clone differed from all other clones and also from the stock culture from which the generation of the clone series started. Differences were found not only in the number and position of bands but also in the brightness of some of them. The underlying cause of the lower and higher brightness can be copy number alterations of the corresponding chromosomal segments (segmental ploidy changes). Of all evidence, the genome of 11–1090 is plastic and its structure frequently changes during the vegetative propagation of its cells. These changes may affect the intensity of pulcherriminic acid biosynthesis (dark red versus brownish pigmentation) but are unlikely to be involved in the reversible ON-OFF switching. Collectively, the phenotypic segregations in the cultures of the ex-type strains of 6 former species recently merged in M. pulcherrima and the 18 generations of clones of the ex-type strain of the former species M. shanxiensis demonstrate that the biosynthesis of pulcherriminic acid is an unstable property and can be stochastically switched on and off during the propagation of cells in a manner reminiscent of the epigenetic switching of gene activities near heterochromatic regions in other yeast species. Once switched on, the intensity of the biosynthesis seems to depend on other factors, for example on which PUL cluster is active in the heterozygous genome and on structural changes in the genome. Since turning off the synthesis of pulcherriminic acid (and pigmentation) is associated with the cessation of the antagonistic effect, it is advisable to test the stability of pigmentation when assessing the antagonistic power of strains. Materials and methods Strains The Metschnikowia strains involved in the investigation of the changes of pulcherrimin production were: 11–578 (CBS 5833) M. pulcherrimaT, 11–579 (CBS 8853) M. pulcherrima (type strain of former species M. fructicola), 11–1088 (CBS 10357) M. pulcherrima (type strain of former species M. sinensis), 11–1089 (CBS 10358) M. pulcherrima (type strain of former species M. zizyphicola), 11–1090 (CBS 10359) M. pulcherrima (type strain of former species M. shanxiensis), and 11–1120 (CBS 10809) M. pulcherrima (type strain of former species M. andauensis). In the antagonism tests, Debaryomyces aff. hansenii Vm7 and Zygosaccharomyces aff. siamensis11–210650 were used as indicator organisms. Strains were maintained on YEA (0.5% yeast extract, 2% glucose and 2% agar) plates. To visualise pulcherrimin production, this medium was supplemented with 0.04 mg/ml FeCl3. At this concentration, all pulcherriminic acid is converted into pulcherrimin within the cells. Since no pulcherriminic acid secretion takes place, the borderlines between the pigment-producing and pigment-free parts of the colonies are sharp. For DNA preparation, cells were cultured in YEL (YEA without agar) broth. The antagonism tests were carried out on EMMA (Edinburgh Minimal Medium51), plates. Test of strains for segregation of pulcherrimin production Pulcherrimin production was examined on YEA plates supplemented with FeCl3. Colonies producing pulcherrimin turn red on this medium. The higher the amount of pulcherrimin produced in the cells, the more intense the pigmentation of the colony. The isolates were inoculated on these plates in three ways: by spreading of cells to obtain individual colonies, by streaking (smearing) cells with a loop and by dropping 10 µl of dense suspension of cells (106 cells/ml). Growth and pigment formation were monitored at 20 °C for four weeks. Production and phenotypic characterisation of 18 generations of (single-cell) clones Cells of the stock culture maintained in the laboratory were suspended in sterile water, and samples of various dilutions of the suspension were spread onto YEA plates supplemented with FeCl3 to obtain individual detached colonies. It was assumed that each colony arose from a single cell. Therefore each colony was considered a clone of vegetatively propagating cells. The pigmentation and the morphology of the colonies (clones) were recorded after two weeks of incubation at 20 °C. Representatives of colonies differing in pigmentation intensity or in patterns of differently coloured sectors or patches (papillae) were isolated and spread onto fresh plates to obtain a new generation of single-cell colonies (clones). The procedure was repeated until the 18th generation was obtained. RAPD-PCR DNA was isolated from overnight cultures grown in YEL broth and the isolated DNA was used for amplification of chromosomal segments with the primers RAPD24 (5‘-GCGTGACTTG-3‘)52, RAPD1283 (5‘-GCGATCCCC A-3‘)53, RF2 (5‘-CGGCCCCTG T-3‘)54, R5 (5′-AACGCGCAAC-3′)55and TtNRSc (5‘-GCTTCTATGGCCAAGTTG-3‘)56 and GoTaq DNA polymerase. The PCR reactions were performed with the following programme: 94 °C for 5 min, 30 × (94 °C for 50s, 38 °C for 50s, 72 °C for 50s), 72 °C for 5 min. The amplified fragments were separated in 1.2% (w/v) agarose gels (1 x TBE buffer) stained with ethidium-bromide. Amplification, cloning, sequencing and translation of PUL2 To find a segment suitable for the examination of the diversity of PUL2, the sequences of the PUL2 genes of the M. pulcherrima strains for which genome sequences (GCA_000317355.2; GCA_003017285.1; GCA_003123635.1; GCA_004217705.1; GCA_009746055.1; GCA_009932455.1; GCA_014905795.1; GCA_029606385.1; GCA_030583425.1; GCA_937857175.1) are available in the NCBI Genome database were aligned. In the MAFFT alignment, a centrally located 411 nt-long segment of the genes proved to be highly variable. This segment (covering the region 827–1267) of PUL2(total gene length: 1440 nucleotides) was amplified with the primers PUL2F (GATCCAGAGCCTCGACGAGATTTTG) and PUL2R (Tgtcacactcgacatgagaatgtctctgt). The PCR product was used for direct sequencing in both directions with the same primers or for random cloning of individual DNA molecules into the pGEM-T Easy Vector (Promega, Madison, United States), following the manufacturer’s instructions. Upon transformation, bacterial colonies were isolated randomly and plasmids were extracted from the isolates. The inserts of the plasmids were sequenced in both directions with the primer pairs used for amplification. The GenBank accession numbers of the sequences are listed in Table. 1 S. The nucleotide sequences were aligned with the MAFFT version 7 multiple alignment algorithm57 available at https://www.ebi.ac.uk/jdispatcher/msa/mafft and translated using the NCBI ORF finder (https://www.ncbi.nlm.nih.gov/orffinder/). Antagonism test The isolates were tested for antimicrobial antagonistic activity by the “Yeast on indicator (YOI)” method58. Briefly, the agar test plate was flooded with a dense suspension (107 cells/ml) of cells of the indicator organism to obtain a continuous layer of cells on the surface of the medium. After removing the rest of the suspension and drying the surface of the plate, loopful amounts of Metschnikowia cultures were smeared on the plate to form spots of ~ 5 mm in diameter. The growth of the spots and the lawn of the indicator around them was monitored at 20 °C for 7 days. If the isolate inhibited or reduced the growth of the lawn, a clear inhibition zone appeared in the lawn around its colony. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements The authors are grateful to Szilvia Struba-Hegedűs and Anita Bordán-Kovács for expert technical assistance. Author contributions M.S. designed the research; isolated the clones, sequenced PUL2, analysed all experimental data and wrote the paper; K.C. conducted the RAPD-PCR experiments. Funding This work was supported by the National Research, Development and Innovation Fund Office, Hungary (2020 − 1.1.2-PIACI-KFI-2020-00130; 2023 − 1.1.1-PIACI_FÓKUSZ-2024-00037) and by the University of Debrecen Programme for Scientific Publication. Open access funding was provided by the University of Debrecen. Data availability All data generated or analysed during this study are included in this published article and its supplementary files. The DNA sequences were deposited in GenBank and the accession numbers are available in Supplementary Table 1 S. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement All data generated or analysed during this study are included in this published article and its supplementary files. The DNA sequences were deposited in GenBank and the accession numbers are available in Supplementary Table 1 S.